Biochimica et Biophysica Acta 1418 (1999) 85^96
Role of the intracellular domain of the L subunit in Na,K pump function Hugues Abriel, Udo Hasler, Ka«thi Geering, Jean-Daniel Horisberger * Institute of Pharmacology and Toxicology, School of Medicine, University of Lausanne, Bugnon 27, CH-1005 Lausanne, Switzerland Received 27 October 1998; received in revised form 29 January 1999; accepted 5 February 1999
Abstract The catalytic K subunit of the (Na,K)- and (H,K)-ATPases needs to be coexpressed with a L subunit in order to produce cation transport activity. Although the isoform of the L subunit is known to influence the functional characteristics of the Na,K pump, the role of the different domains of the L subunit is not fully understood. We have studied the function of a Na,K pump resulting from the expression of a wild-type K subunit with a N-terminally truncated mutant of the L subunit using the two-electrode voltage clamp and the cut-open oocyte techniques. While the maximal activity, measured as the K activated outward current, was not significantly altered, the L N-terminal truncation induced an ouabain-sensitive conductance in the absence of extracellular K . The voltage dependence of the ouabain-sensitive charge distribution indicated that in the Na/Na exchange conditions, the E1-E2 conformation equilibrium was shifted towards the E2 conformation, a change resulting from alteration of both the forward and the backward reaction rate. Removal of the intracellular domain of the L subunit modifies several aspects of the whole enzyme function by a mechanism that must imply the state of the extracellular and/or transmembrane parts of the K/L subunit complex. ß 1999 Elsevier Science B.V. All rights reserved. Keywords: Sodium pump; L Subunit; Structure-function relationship; Transient current; Cut-open oocyte
1. Introduction The Na,K pump is a ubiquitous membrane bound ATPase which carries sodium ions out and potassium ions into the cell and thereby maintains the Na and K gradients across the membrane of all animal cells. These ionic gradients are involved in a number of important cell functions like membrane excitability, cell volume regulation, and transepithelial secretion and reabsorption [1]. The Na,K pump is composed of an K and a L subunit. In the kidney a
* Corresponding author. Fax: +41 (21) 6925355; E-mail:
[email protected]
Q subunit seems also to be associated [2,3]. The K subunit is known as the `catalytic' subunit because ATP binding and hydrolysis are performed by an intracellular domain of this protein [1]. The L subunit is known to be essential for biosynthesis, maturation and migration of the functional Na,K pump from the endoplasmic reticulum to the cell surface [4]. Moreover, this L subunit has been shown to be involved in functional aspects of the active Na,K pump, namely in the apparent a¤nity for extracellular K [5^7]. According to the Albers-Post kinetic scheme, the Na,K pump exists alternately in two major conformations during the Na,K transport [8,9]. The E1 conformation has a high a¤nity for intracellular
0005-2736 / 99 / $ ^ see front matter ß 1999 Elsevier Science B.V. All rights reserved. PII: S 0 0 0 5 - 2 7 3 6 ( 9 9 ) 0 0 0 2 5 - 5
BBAMEM 77584 2-4-99
86
H. Abriel et al. / Biochimica et Biophysica Acta 1418 (1999) 85^96
Na which can access to its intracellular cation binding sites and the E2 conformation, which exposes high a¤nity sites for extracellular K . The activity of the Na,K pump is electrogenic since it moves 3 Na out and 2 K into the cell during a full transport cycle. Experimental evidence suggests that the main charge translocation occurs during the transport and release to the outside of sodium ions [10^ 12]. This Na -dependent charge translocation has allowed the study of the kinetics of single steps in the Na transporting limb of the cycle, i.e. when the Na,K pump activity is studied in the Na /Na exchange mode. Indeed, this could be done by measurements of ouabain-sensitive pre-steady-state currents following fast voltage perturbation in myocardial cells [13,14], in Xenopus oocytes with endogenous pump [15,16] and arti¢cially expressed wild-type and mutant Na,K pump subunits [17]. It has been shown that truncation of the N-terminal 34 amino acids, comprising the whole cytoplasmic domain of the L subunit, induced a signi¢cant decrease in the apparent a¤nity of the Na,K pump for extracellular K [6]. In a recent detailed work [18], it has been proposed that the N terminus is not directly involved in the K e¡ect but that truncation perturbs the ectodomain and/or the transmembrane domain of the L subunit. The mechanism of this structural perturbation remains to be explained. Functionally, it has been shown that the apparent a¤nity for extracellular K was decreased by a factor 2^8, depending on the presence of extracellular Na [18]. Moreover, the apparent a¤nity for intracellular Na was also decreased by a factor of about 2 in Na,K pumps containing the N-terminally truncated L subunit. In order to understand how the removal of an intracellular domain of the `accessory', non-catalytic subunit could have such a large e¡ect on the function of the Na,K pump, and in particular on the apparent a¤nity for extracellular K , we have characterized in more detail several functional aspects of Na,K pumps containing a wild-type or a N-terminally truncated L subunit. Using the classical two-electrode voltage-clamp technique and a modi¢ed cut-open oocyte technique, we obtained evidence indicating that the L N-terminal truncation modi¢es several functional properties of the Na,K pump, in addition to the apparent a¤nities for cations, in particular the E1-E2 conformation equilibrium and the
barrier property. In accordance with recent molecular analysis [18], our results suggest that truncation of the N-terminal intracellular domain of the L subunit results in a structural modi¢cation of the transmembrane and extracellular parts of the whole enzyme. 2. Materials and methods 2.1. Generation of the N-terminally truncated mutants of the L subunit and expression of the Na,K pump in Xenopus oocytes The production of a mutant of the (Na,K)-ATPase L isoform with a 34-amino acid N-terminal truncation (t34L mutant) has been previously reported [6]. Stage V/VI Xenopus laevis oocytes were obtained from ovarian tissue of females which had been anesthetized by immersion in MS 222 (2 g/l; Sandoz, Basel, Switzerland). The oocytes were defolliculated as previously described [19]. Wild-type (WT) and mutant Xenopus L subunit cRNA (1 ng) was injected with wild-type Bufo K subunit cRNA (7 ng) in a total volume of 50 nl water, into oocytes as already described [19]. The oocytes were kept in a modi¢ed Barth's solution and electrophysiological measurements were performed 3 days later. The oocytes were Na loaded by a 2 h incubation in a K -free and Ca2 -free solution containing 90 mM Na and 0.5 mM EGTA. They were then kept for at least 2 h in a K -free solution with 0.4 mM Ca2 and 200 nM ouabain [20]. These conditions allowed a selective study of the exogenous Na,K pumps composed of Bufo K1 (ouabain resistant) and Xenopus L1 subunits, in the presence of inhibited, endogenous ouabainsensitive, Xenopus Na,K pumps [5]. 2.2. Measurements of steady-state currents with the two-electrode voltage-clamp technique We used the two-electrode voltage-clamp technique to measure the steady-state Na,K pump currents, as described earlier [5]. Voltage and current were recorded with a TEV-200 voltage-clamp (Dagan, Minneapolis, MN, USA). In each oocyte the Na,K pump current, i.e. the outward current activated by addition of 10 mM K to an initially K -free solu-
BBAMEM 77584 2-4-99
H. Abriel et al. / Biochimica et Biophysica Acta 1418 (1999) 85^96
tion, was measured at 350 mV. This value has been shown to be a good estimate of the number of Na,K pumps at the oocyte surface [21] and was used to normalize the current values in the di¡erent experimental conditions. The voltage dependence of the ouabain-sensitive currents was determined by measuring the current during a series of nine 250 ms equal voltage steps (3130 to +30 mV or 3170 to +70 mV). This was done before and after the addition of 2 mM ouabain. When the e¡ect of ouabain was studied in solutions with di¡erent pH (7.4 and 6.0), measurements were ¢rst performed in each oocyte under the two conditions in the absence of ouabain. Two mM of ouabain was then added and the same series of measurements was repeated 2 min later. 2.3. Measurements of pre-steady-state currents with the cut-open oocyte technique In order to obtain a voltage clamp with high time resolution for the investigation of short pre-steadystate current due to transient charge movement by the pump, we used a modi¢ed cut-open oocyte technique. This technique, originally developed by Taglialatela et al. [22], has already been used for the study of the pre-steady-state current of the (Na,K)ATPase by Holmgren and Rakowski [16]. Brie£y, a Xenopus oocyte was mounted between two compartments with the vegetal pole upwards. The superior pole of the oocyte was in contact with the upper bath through a hole of 0.9 mm in diameter. The middle (guard) bath served to provide electrical isolation between the upper (extracellular) and lower (intracellular) compartments. This was obtained through independent voltage clamping of the middle bath at the same electrical potential as the upper one. The lower pole of the oocyte was impaled with a glass microelectrode which was simultaneously used as an internal perfusion pipette and a voltage-recording electrode. This modi¢cation to the original set-up was ¢rst described and successfully used by Costa et al. [23]. The resistance of the electrode, when ¢lled with the intracellular solution described below, was about 0.2^0.7 M6. The external compartment was superfused by gravity (£ow rate approx. 6 ml/min) with an extracellular Na -containing solution (see below). However, the pre-steady-state currents were measured under no £ow conditions in order to reduce
87
the noise associated with the extracellular £uid perfusion. The internal solution was perfused by means of a precision syringe pump (Infors, Basel, Switzerland) at a £ow rate of 1 Wl/min. The voltage clamp was performed using a Dagan cut-open oocyte apparatus (Dagan, Minneapolis, MN, USA; Model CA-1 High Performance Oocyte Clamp). Data acquisition and analysis were performed using a TL1 DMA digital converter system and the pClamp software package (Axon Instruments, Foster City, CA, USA; version 5.5). The holding potential was 350 mV. Voltage pulses of 30 ms were made from the holding potential over the range of 3190 to +90 mV in increments of 35 mV. Twenty runs of recordings were averaged. The current signal was ¢ltered at 1 or 2 kHz using a four-pole Bessel ¢lter. The ouabain-sensitive transient currents were obtained by subtraction of the currents measured 2 min after the superfusion of the same solution but containing 2 mM ouabain from that measured before. 2.4. Determination of Na,K pump turnover rate The turnover rate of the wild-type and mutated Na,K pump was determined by dividing the mean K -activated outward current (in charges per second) by the mean number of ouabain binding sites determined in a group of oocytes from the same batch, as previously described [21]. These experiments were performed using the K subunit from X. laevis in order to allow the measurements of high a¤nity ouabain binding sites (ouabain binding measurements are technically di¤cult and prone to large errors with the ouabain-resistant Bufo K1 isoform). Brie£y, Na -loaded oocytes were exposed for 30 min to a K -free solution containing 0.67 WM cold ouabain and 0.33 WM 3 H-ouabain (Amersham). Oocytes were then washed and individually counted after solubilization in 100 Wl of 5% SDS and addition of 2 ml of emulsi¢er scintillator plus scintillation liquid (Packard). 2.5. Solutions and drugs The compositions of the solutions used for the presteady-state currents and pump turnover measurements were as follows: extracellular Na -containing solution (mM): Na 92.4, Mg2 0.82, Ba2 5, Ca2
BBAMEM 77584 2-4-99
88
H. Abriel et al. / Biochimica et Biophysica Acta 1418 (1999) 85^96
0.41, tetraethylammonium (TEA) 10, Cl3 22.5, HCO3 3 2.4, gluconate 80 and HEPES 10, pH 7.4; solution for the intracellular perfusion: Na 50, K 50, Mg2 0.82, gluconate 100, Cl3 1.64, Trisadenosinetriphosphate (ATP) 5, EGTA 2, pH 7.35. For the pH-dependent ouabain-sensitive current experiments the `control' K -free solution contained Ba2 5, TEA 10, Mg2 0.82, Ca2 0.41, (N-morpholino)propanesulfonic acid (MOPS) 53, Cl3 22, sucrose 170. The pH was adjusted to 7.4 with N-methyl-D-glucamine (NMDG). Ten mM K was added as K-gluconate (osmolarity was not compensated for) for activation of the Na,K pump. Ouabain was obtained from Sigma (St. Louis, MO, USA). All the experiments were performed at room temperature (22^27³C). Results are expressed as mean þ S.E.M., n = number of tested oocytes if not stated otherwise. The statistical signi¢cance of di¡erences between means was estimated using bilateral Student's t-test for unpaired data. 3. Results 3.1. Steady-state current in the wild-type and L N-terminally truncated Na,K pump
Fig. 1. Steady-state ouabain-sensitive current-voltage relationships in the presence of 10 mM K and in K -free solution. (A) Forward Na,K pump currents were activated with 10 mM K and ouabain-sensitive currents (Ip ) were determined by subtraction of current-voltage relationships before and 2 min after addition of 2 mM ouabain in both groups: wild-type (b, n = 8) and L N-terminally truncated Na,K pump (a, n = 4). At 3170 mV, Ip was signi¢cantly smaller in the L N-terminally truncated Na,K pumps as compared to the wild-type. (B) In a K - and Na -free solution at pH 6.0, a larger inward ouabain-sensitive current (Iou inw) was observed with the L N-terminally truncated Na,K pump (a, n = 15) compared to the wild-type pump (b, n = 14). In the presence of extracellular Na and at pH 7.4, Iou inw was very small in wild-type (F, n = 5) but clearly larger in the L N-terminally truncated Na,K pump (E, n = 5). The ouabain-sensitive inward currents were normalized to the 10 mM K -activated Na,K pump current at 350 mV.
As already observed by Hasler et al. [18], using the two-electrode voltage-clamp technique, the mean ouabain-sensitive steady-state Na,K pump current (Ip ) (activated with 10 mM K at 350 mV) was similar for the wild-type (222 þ 19 nA, n = 8) and L N-terminally truncated (219 þ 29 nA, n = 4) Na,K pumps. Fig. 1A shows the current-voltage relationships of the ouabain-sensitive current of both Na,K pumps. At large negative membrane potential values ( 6 3150 mV), Ip tended to be smaller in the L Nterminally truncated compared to the wild-type Na,K pumps. This di¡erence was statistically signi¢cant (P 6 0.05) at 3170 mV: 112 þ 15 nA (n = 8) for the wild-type Na,K pump versus 29 þ 32 nA (n = 4) for the L N-terminally truncated Na,K pump. 3.2. Ouabain-sensitive inward-rectifying current (Iou inw) To further characterize the biophysical properties of the mutant pump, we studied the ouabain-sensi-
BBAMEM 77584 2-4-99
H. Abriel et al. / Biochimica et Biophysica Acta 1418 (1999) 85^96
tive inward current (Iou inw) which can be measured with extracellular Na - and K -free solutions [17,24,25]. The amplitude of this current is larger with extracellular solutions of low pH and it has been proposed to be an inward proton current £owing through the E2 conformation of the Na,K pump [25]. With an extracellular solution of pH 6.0, which increases the Iou inw, the normalized inward current was about twice larger (at 3130 mV) for the L Nterminally truncated as compared with that for the wild-type Na,K pump (Fig. 1B and Table 1). The amplitude of Iou inw was smaller in both groups at neutral extracellular pH, but it was still about 3 times larger in the L N-terminally truncated than in the wild-type Na,K pumps (Table 1). We further measured the ouabain-sensitive current in the presence of extracellular sodium (92 mM) at pH 7.4, a condition in which almost no ouabain-sensitive conductance can be detected [5]. We could still observe a sizeable Iou inw in the L N-terminally truncated mutant Na,K pump which, at 3130 mV, reached about one third of the forward Na,K pump (K -activated) current at 350 mV (Fig. 1B and Table 1). With the wild-type pump this current was reduced to less than one tenth of the forward current (Table 1). 3.3. Transient currents in the Na+/Na+ exchange mode In the absence of extracellular K and with high concentrations of Na on both sides of the membrane, the (Na +K )-ATPase functions in the socalled Na /Na exchange mode [26]. For this exchange, we can use the simpli¢ed reaction scheme [13]:
89
Nai E13E2 Nao : It has been shown that during this exchange, charges are translocated across the membrane inducing measurable transient (pre-steady-state) currents when the membrane potential is rapidly changed [13,15^17]. A simple model has been proposed to explain this observation: the transport of three sodium ions out of the cell and the associated E1 to E2 conformational change result in the translocation of one net charge across the membrane electrical ¢eld. Thus, the time course of the transient current relaxation re£ects the redistribution of the E1 and E2 states towards the new equilibrium imposed by the membrane potential. Because no transient current could be detected in the K /K exchange conditions [27], a simple model has been proposed in which the translocation of two potassium ions is electroneutral because during the translocation, these ions are associated with two negative charges of the protein while a transfer of one net charge occurs when three sodium ions are transported across the membrane, two of the three sodium ions being associated with the protein negative charges. Further results have shown that the charge translocation mechanism may be more complicated with several components of di¡erent time course [14,28,29] and that the transfer of K may also be accompanied by charge translocation [30]. In our experimental conditions, using the cut-open oocyte technique, the fast applied voltage steps caused large transient currents mainly due to the capacitance of the oocyte membrane (Fig. 2A,B). The pre-steadystate currents due to the Na,K pump were obtained by subtracting the current values before and after ouabain application. Fig. 2A and B show original
Table 1 Ouabain-sensitive inward current in K -free solutions as a function of extracellular pH and Na concentration Iou inw at 3130 mV (normalized)
pH 6.0, Na 0 mM pH 7.4, Na 0 mM pH 7.4, Na 92 mM
L wild-type
L N-terminally truncated
0.97 þ 0.18 n = 12 0.44 þ 0.07 n = 10 0.08 þ 0.03 n=5
1.89 þ 0.28* n = 15 1.23 þ 0.26** n=8 0.38 þ 0.05* n=5
All values were normalized to the maximal K activated outward current measured immediately before in the same oocyte. * and ** indicate mean values signi¢cantly di¡erent from that of the wild-type group (P 6 0.05 and 0.01 respectively); n, number of observations.
BBAMEM 77584 2-4-99
90
H. Abriel et al. / Biochimica et Biophysica Acta 1418 (1999) 85^96
Fig. 2. Original recordings of pre-steady-state currents obtained using the cut-open oocyte technique. (A,B) Original recordings of the total current across the oocyte membrane patch (0.9 mm of diameter) recorded before (A) and 2 min after (B) addition of 2 mM ouabain in an oocyte expressing the wild-type Na,K pump. The inset in A shows a recording of the corresponding voltage traces during the series of 30 ms voltage steps from 3190 to +90 mV; the calibration bars are 10 ms and 100 mV. The ouabain-sensitive currents (C) were obtained by subtraction of currents in A and B. Panel D shows an example of ouabain-sensitive currents in an oocyte expressing the L N-terminally truncated Na,K pump; the presence of an inward steady-state current can be noticed. For both types of pump, the `on' and `o¡' current relaxations starting 2 ms after the membrane voltage change could be well ¢tted to a monoexponential decrease as illustrated in insets in C and D. Calibration bars are 10 ms and 1 WA for both insets.
recordings of the voltage and current traces before and after addition of 2 mM of ouabain to the extracellular solution. Using the cut-open oocyte technique, a stable membrane voltage was obtained after about 1 ms (Fig. 2A, inset). The ouabain-sensitive current component is easily seen when the tail currents in both conditions are compared. Fig. 2C and D show the presence of an ouabain-sensitive transient current observed in oocyte expressing wild-type and L N-terminally truncated Na,K pumps. About 2 ms after the command voltage step, the transient
current decreased with a relaxation rate which could be ¢tted with a mono-exponential function (insets in Fig. 2C,D). The current in the ¢rst 2 ms was too variable between the individual preparations and was not considered for analysis. We therefore restricted our measurement to the slow component of the ouabain-sensitive pre-steady-state current, as also done in previous work [17]. Similarly to what has been described above about the results obtained using the two-electrode voltageclamp technique (Fig. 1B), we also observed using
BBAMEM 77584 2-4-99
H. Abriel et al. / Biochimica et Biophysica Acta 1418 (1999) 85^96
91
the cut-open oocyte technique that very little steady state ouabain-sensitive current could be detected in the wild-type Na,K pump (Fig. 2C) while we observed a sizeable ouabain-sensitive steady-state current in the L N-terminally truncated mutant (Fig. 2D).
the wild-type and mutant Na,K pump. For the wildtype Na,K pump the relationship was sigmoidal and could be well ¢tted with the Boltzmann equation:
3.4. Ouabain-sensitive charge translocation
where F is the Faraday constant, R is the Boltzmann constant, T is the temperature, Qmax is the maximal displaceable charge, Vq is the midpoint potential, zq a steepness factor and Vm the membrane voltage. The values of the parameters of the Boltzmann equation (see legend of Fig. 4) obtained from the mean data of the wild-type Na,K pump group were in reasonable agreement with those obtained by us and others using various techniques considering that, in these experiments, only about one third of the total membrane surface was studied [13,15^17]. The shape of the Qou /Vm relationship was di¡erent for the L N-terminally truncated mutant Na,K pumps. As shown in Figs. 3B and 4A, the curve did not present the characteristic sigmoidal shape as it did not saturate at negative potential values in the explored voltage range. Attempts to ¢t the parameters of the Boltzmann equation to this curve yielded a much reduced steepness factor (about 3^4 times smaller than for wild-type pumps, see Fig. 4A),
The quantity of charge moved during the `on' (Qon ) and the `o¡' (Qoff ) phases of the voltage step were, in a ¢rst analysis, calculated using the integration routine of the pClamp program between 0 and 20 ms after the start of the voltage change. Because of the presence of an ouabain-sensitive steady-state current with the L N-terminally truncated mutant (see above), the amount of charge due to the steady-state component was subtracted from the total current integration to estimate true Qon , i.e. the charge component due to the transient current only. As shown in Fig. 3A (wild-type) and B (L N-terminally truncated mutant) the absolute values of Qon and Qoff were similar. The mean values of the ouabain-sensitive charge displacement (Qou ) were calculated from 3Qoff to avoid errors due to the steadystate current in L N-terminally truncated Na,K pump. Fig. 4A shows the Qou /Vm relationships for
Qou Qmin Qmax =
1 expzq
V q 3V m
F =RT
1
Fig. 3. Ouabain-sensitive charge translocation Qon and Qoff with exogenous expressed wild-type and L N-terminally truncated pumps in typical experiments using the cut-open oocyte technique. (A) The charge translocation Qon (b) and 3Qoff (a), calculated as time integral of the pre-steady-state current vs. the membrane potential, is represented for the wild-type pump. There was no di¡erence in absolute values between the Qon and Qoff . The relationship shows saturation at high negative and positive potentials. (B) For the L Nterminally truncated mutant Na,K pump, the Qon (b) were identical to the 3Qoff (a) provided that the amount of charge due to the steady-state current component was subtracted from Qon (see text). Compared to the wild-type relationship, the Q-values showed no tendency to saturation at high negative potentials.
BBAMEM 77584 2-4-99
92
H. Abriel et al. / Biochimica et Biophysica Acta 1418 (1999) 85^96
Fig. 4. Ouabain-sensitive charge translocation (Qou ) as a function of membrane voltage using two modes of Qou estimation. (A) The charge translocation, obtained as time integral of the recorded current, is represented as a function of membrane voltage. The curves are drawn using the best ¢tting parameters from Eq. 1 on mean values. For the wild-type Na,K pump (b, n = 17) and L N-terminally truncated mutant (a, n = 19), the values of Qmax (maximal translocated charge) were 6.6 and 12.6 nC, respectively, the values of zq (steepness factor) 0.71 and 0.19, respectively and the values of Vq (midpoint potential), 362 and 3192 mV, respectively. (B) Slow component of the ouabain-sensitive charge displacement estimated from the monoexponential current relaxation between 2 and 20 ms (see text). Curves are drawn using the best ¢tting parameters from the Boltzmann equation on mean values. For the wild-type group (b, n = 17) and L N-terminally truncated mutant (a, n = 17), the Qmax were 9.0 and 12.9 nC, the zq 0.68 and 0.25 and the Vq 368 and 3142 mV, respectively.
a more negative value of midpoint potential for the mutant compared to that of the wild-type Na,K pumps and a larger maximal displaceable charge (Qmax ). The mode of estimation of the ouabain-sensitive displaced charge using current integration includes in theory all the components (fast and slow) of the charge transfer. However, due to the uncertainty and the variability of the very initial currents (¢rst 2 ms), this mode of calculation may be prone to errors. We therefore also analyzed speci¢cally the charge translocation due to the slow component by calculating the integral over time of the ouabain-sensitive current, assuming an exponential decrease in this current with a rate constant Kon : I I
0 exp
3K on t
2
Using the values of Kon obtained from the best ¢tting monoexponential to the current decay between 2 and 20 ms after the voltage step and the amplitude of the current at time 2 ms current values, we calculated Qou as the integral over time of Eq. 2 extrapolated from t = 0 to r. These Qou values are presented on Fig. 4B. The best ¢tting parameters of the Boltzmann equation were obtained for the set of Qou (V) values obtained in each oocyte. The mean values of these parameters were respectively for the wild-type
(n = 17) and the mutant (n = 17) groups: Qmax 9.1 þ 1.2 and 14.5 þ 2.2 nC, P 6 0.05, zq 0.73 þ 0.07 and 0.28 þ 0.03, P 6 = 0.001, and Vq 369 þ 2 and 3172 þ 28 mV, P 6 0.001. It has to be noted that because the range of our data covered only a part of the sigmoidal curve in the case of the mutant Na,K pump, the values of these parameters should be taken with care. But, despite this reservation, two di¡erences are striking: the low slope of the curve indicates that the voltage dependence of the charge translocation is reduced and the large shift to the left of the Qou /Vm relationship suggests that the mutant Na,K pump is strongly shifted towards the E2 conformation at physiological membrane potentials. 3.5. The relaxation rates of the ouabain-sensitive currents The relaxation rates of the `on' tail current (kon ) and `o¡' tail current (koff ) were determined by ¢tting a single exponential equation to the current values starting 2 ms after the potential change. Over this time frame, the ¢t was good (see Fig. 2C,D, insets), and could not be signi¢cantly improved with a second order exponential equation. The mean values of
BBAMEM 77584 2-4-99
H. Abriel et al. / Biochimica et Biophysica Acta 1418 (1999) 85^96
93
followed by a test pulse to +50 mV. As shown in Table 2, the mean Kon obtained from 3150 to +50 mV voltage jumps was about twice faster with the L N-terminally truncated than with the wild-type Na,K pumps. 3.6. Turnover rates of the wild-type and L N-terminally truncated Na,K pump
Fig. 5. Voltage dependence of the relaxation rate constants (k) of the pre-steady-state ouabain-sensitive current. The mean `on' relaxation rate (at the beginning of the voltage step) is shown as a function of voltage. Wild-type (n = 22) and L N-terminally truncated (n = 18) values are represented by ¢lled and open circles, respectively. The mean `o¡' relaxation rate (upon return to 350 mV at the end of the voltage step) is given by the large symbols at 350 mV. Note the strong voltage dependence of the kon values for the wild-type Na,K pump at high negative membrane potentials compared with the low voltage dependence of the L N-terminally truncated mutant Na,K pump. * and ** indicate values signi¢cantly di¡erent from that of the wild-type group (P 6 0.05 and 0.01 respectively).
kon for both groups expressed as a function of voltage are shown in Fig. 5. The mean values of koff , which were similar for all voltage steps, are indicated by the large symbols at the 350 mV point. For the wild-type Na,K pump, kon showed a clear voltage dependence with rates larger at negative potentials. In comparison, the kon of the L N-terminally truncated Na,K pump was almost independent of voltage. At negative potential values, the relaxation rates were signi¢cantly smaller in the mutant group and at positive values, they were signi¢cantly higher (Fig. 5 and Table 2). Because of the large shift of the Qou / Vm curve towards negative potentials, the absolute value of charge translocation after voltage step from 350 mV to more positive potentials was small in the L mutant Na,K pumps, resulting in a small signal to noise ratio. In order to examine the voltage dependence of the relaxation rate constant with a large signal and obtain more precise measurements, we used the following pulse protocol: from the holding potential, a pre-pulse of 35 ms at 3150 mV was
The estimation of Qmax was larger with the L Nterminally truncated Na,K pump than with the wildtype Na,K pump (Fig. 4), while there was no signi¢cant di¡erence in the steady-state Na,K pump current [18]. This observation could be explained by a slower turnover rate of the L N-terminally truncated Na,K pump if expression of these mutants resulted in a higher density of Na,K pumps. We therefore determined the turnover rate in the two groups by dividing the mean steady-state pump current by the mean number of ouabain binding sites for a given batch of oocytes. For these experiments, we injected the Xenopus K subunit, instead of the ouabain-resistant Bufo subunit, in order to be able to perform the ouabain binding experiment. There was no signi¢cant di¡erence between the two groups either for the pump current (wild-type 436 þ 61 nA and L N-terminally truncated 327 þ 88 nA, n = 3 batches of oocytes with 6^10 oocytes per batch) or for the number of ouabain binding sites (wild-type 48.5 þ 3.9U109 sites/oocyte and L N-terminally truncated 47.4 þ 5.6U109 sites/oocyte). The calculated turnover rates for the wild-type and L N-terminally truncated Na,K pumps were 55 þ 3 and 44 þ 12 charges/s, respectively. For the endogenous Xenopus Table 2 Relaxation rate constant (kon ) of the pre-steady-state ouabainsensitive current kon (s31 )
L wild-type
L N-terminally truncated
from 350 to 3190 mV
1022 þ 50 n = 22 510 þ 46 n = 22 313 þ 14 n = 11
648 þ 47** n = 18 704 þ 73* n = 18 573 þ 21** n=9
from 350 to +90 mV from 3150 to +50 mV
* and ** indicate mean values signi¢cantly di¡erent from that of the wild-type group (P 6 0.05 and 0.01 respectively).
BBAMEM 77584 2-4-99
94
H. Abriel et al. / Biochimica et Biophysica Acta 1418 (1999) 85^96
pump, i.e. in non-injected oocytes, the current was about only one tenth of the exogenous one, but the turnover was also in the same range (59 þ 12 charges/ s). 4. Discussion Among the large family of the P-type ATPases, the (Na,K)-ATPase together with the gastric and nongastric types of (H,K)-ATPase form a distinct group of ion pumps. All known members of this group carry out the transport of potassium ions inside the cell in exchange for the extrusion of Na or H out of the cell. One common and speci¢c characteristic of this subgroup of P-ATPase is the presence of a L subunit associated with the main `catalytic' K subunit. The presence of this L subunit seems to be absolutely required for the expression of a cation transport function by this group of proteins (for a review see [7]), although the K subunit alone may be able to perform ATP hydrolysis [31]. While the role of the L subunit in the maturation process has been well documented [32], its role in the mature enzyme expressed at the cell surface is not as well understood. Functional studies in several expression systems have shown that the type of the L subunit that is associated with the K subunit in£uences the function of the whole enzyme (for review see [7]), in particular its apparent a¤nity for extracellular K [5]. Although the three main domains (intracellular N-terminal, transmembrane and extracellular C-terminal domains) of the L subunit have been implicated in this e¡ect, the mechanism of the K/L subunit interaction is still poorly understood [6,32]. Hasler et al. [18] have recently shown that a truncation of the whole N-terminal domain of the L subunit induces a large reduction of the apparent a¤nity for extracellular K and a smaller reduction of the apparent a¤nity for intracellular Na . In the present work, we have explored in more detail the e¡ects of the N-terminal truncation of the L subunit on the function of the Na,K pump. As recently reported [18], we found that the L Nterminally truncated Na,K pump was able to associate with the K subunit and to form a fully functional Na,K pump at the oocyte surface. The maximal rate of cation transport estimated from the ratio between
the K -activated Na,K pump current (under Vmax conditions) to the number of ouabain binding sites was similar in wild-type Na,K pumps and pump containing the N-terminally truncated L subunit. The current-voltage relationship of the ouabain-sensitive current was only slightly altered (see Fig. 1A), with a small increase in the voltage dependence of this current at low membrane potentials. These observations suggest that the rate-limiting transport reactions were little altered by the N-terminal truncation of the L subunit. More detailed measurements revealed, however, that the Na,K pump function was nevertheless profoundly modi¢ed. A striking di¡erence between wild-type and L Nterminally truncated Na,K pumps was the presence of an ouabain-sensitive inward current when the mutant was studied in the absence of external K , a condition in which no Na/K exchange can occur. The role of Na,K pumps is to establish and maintain the gradients of Na and K across the plasma membrane. An e¤cient ion pump function requires that little leak of ions occurs, otherwise energy (ATP) would be spent with no gain. Indeed, the normal (Na,K)-ATPase is very `tight'. Under conditions in which no active transport occurs, such as the absence of extracellular K , hardly any ouabain-sensitive ion conductance can be detected [5], except for the pHsensitive conductance when the measurements are performed in the absence of extracellular Na and K [25], i.e. under far from physiological conditions. As shown in Fig. 1B, the Na,K pump containing a L N-terminally truncated subunit showed a signi¢cant `leak', demonstrated by the presence of an ouabainsensitive conductance at a physiological pH and external Na concentration. The amplitude of the pHsensitive conductance was also signi¢cantly increased in the L N-truncated mutant. The steeper I/V curve of the steady-state ouabain-sensitive current observed in the presence of external K suggest that the `leak' conductance is also present in the presence of K . This hypothesis, however, would be di¤cult to demonstrate, because we have no easy way to distinguish which part of the total ouabain-sensitive current is due to the forward Na/K exchange cycle and which part could be due to a `leak' conductance. The transport of three sodium ions out of the cell, and the associated E1 to E2 conformation change, are accompanied by the translocation of one net
BBAMEM 77584 2-4-99
H. Abriel et al. / Biochimica et Biophysica Acta 1418 (1999) 85^96
charge across the membrane electrical ¢eld [13,33]. More precisely, the main charge translocating event seems to be associated with the deocclusion and release of the ¢rst sodium ion to the outside medium [14]. This charge movement can be measured as a pre-steady-state ouabain-sensitive current following rapid membrane voltage perturbations under conditions of Na /Na exchange. With the wild-type Na,K pump, the voltage dependence of the charge translocation can be well described by a Boltzmann function, as shown in several previous reports [13,16,17] and by our results (Fig. 4), with a steepness factor z of about 0.75. In contrast, in the L Nterminally truncated Na,K pump, the voltage dependence was much reduced and no saturation could be observed at membrane potentials as negative as 3190 mV. The best ¢tting parameters of the Boltzmann equation indicated a larger maximal displaceable charge. Since no increase in the density of ouabain binding sites could be detected when the Xenopus K1 subunit was expressed with the N-truncated L subunit (when compared with the wild type L subunit expressed with the same K subunit), these results suggest that the modi¢ed Na,K pump carries out a larger maximal charge displacement upon voltage perturbation. Whether this change is related to a redistribution between the fast and slow components of the charge-translocating events or whether it due to the presence of the movement of additional intrinsic charges in the membrane electrical ¢eld cannot be determined by our present results. In any case, these results have to be regarded with caution since rather large errors could be made in the estimation by extrapolation of the maximal charge displacement because of the low slope of the Qou /Vm relationship. The midpoint potential, which had a value of about 360 mV in wild-type Na,K pump, was displaced toward much more negative potentials in the L N-truncated pumps. This suggests that at 360 mV the L N-truncated Na,K pump is mostly in the E2 conformation, while the wild-type Na,K pump is nearly equally distributed between the E1 and the E2 conformation. This interpretation is supported by analysis of the kinetics of the charge translocation (Fig. 5) which show a slower inward (E2CE1) and a faster outward (E1CE2) charge translocation in the mutant Na,K pump.
95
Taken together, with the earlier published results obtained with the same mutant showing a 2^8-fold reduction in apparent a¤nity for extracellular K (depending on the presence or absence of extracellular Na ) and an approx. 2-fold reduction in apparent a¤nity for intracellular Na [18], our results indicate that the N-terminal truncation of the L subunit produces complex and profound changes in the cation transport function of the Na,K pump, even though these changes have little e¡ect on the Na,K pump activity maximally stimulated by extracellular K . In a previous study, using a simple kinetic model of the Na,K pump transport cycle [17], we have shown that N-terminal truncation of the K subunit induces changes in the apparent a¤nity for extracellular K that could be well explained by a modi¢cation of the rate of the E1CE2 conformation change. Using the same kinetic model we also attempted to reproduce the values for apparent Na and K af¢nities observed in the L N-terminally truncated Na,K pump by applying the values of the rate of conformation change deduced from the charge translocation measurements. However, in the case of the N-terminally truncated L subunit, we were not able to explain the large change in the apparent a¤nity for extracellular K only by modi¢cation of the E1HE2 conformation equilibrium; a modi¢cation of the intrinsic a¤nity for extracellular K was necessary. We conclude that removal of the intracellular Nterminal domain of the L subunit induces signi¢cant changes in the conformation of the whole K/L complex, including modi¢cations of the external K binding site, changes in the E1HE2 conformational equilibrium and formation of a `leaky' E2 conformation. This suggests that the state of the intra-membrane part at least, and possibly also of the extracellular part of the whole K/L complex, is dependent on the integrity of the intracellular N-terminal domain of the L subunit. This conclusion is in good agreement with those of Hasler et al. [18] who suggest that functionally important K/L subunit interactions occur between the extracellular domains of these subunits rather than between the intracellular domains. The absence of the N-terminal domain of the L subunit appears to entail important modi¢cations of the whole K/L complex.
BBAMEM 77584 2-4-99
96
H. Abriel et al. / Biochimica et Biophysica Acta 1418 (1999) 85^96
Acknowledgements This work was supported by the Swiss Fonds National de la Recherche Scienti¢que grant No. 3145867.95 References [1] J.-D. Horisberger, The Na,K-ATPase: Structure-Function Relationship, Landes, Austin, TX, 1994. [2] R.W. Mercer, D. Biemesderfer, D.P. Bliss Jr., J.H. Collins, B.I. Forbush, J. Cell Biol. 121 (1993) 579^586. [3] P. Beguin, X.Y. Wang, D. Firsov, A. Puoti, D. Claeys, J.-D. Horisberger, K. Geering, EMBO J. 16 (1997) 4250^4260. [4] K. Geering, FEBS Lett. 285 (1991) 189^193. [5] F. Jaisser, P. Jaunin, K. Geering, B.C. Rossier, J.-D. Horisberger, J. Gen. Physiol. 103 (1994) 605^623. [6] K. Geering, A. Beggah, P. Good, S. Girardet, S. Roy, D. Schaer, P. Jaunin, J. Cell Biol. 133 (1996) 1193^1204. [7] D.C. Chow, J.G. Forte, J. Exp. Biol. 198 (1995) 1^17. [8] R.W. Albers, Annu. Rev. Biochem. 36 (1967) 727^756. [9] R.L. Post, C. Hegyvary, S. Kume, J. Biol. Chem. 247 (1972) 6530^6540. [10] W. Stu«rmer, R. Bu«hler, H.-J. Apell, P. La«uger, J. Membr. Biol. 121 (1991) 163^176. [11] D.C. Gadsby, R.F. Rakowski, P. De Weer, Science 260 (1993) 100^103. [12] K. Fendler, E. Grell, E. Bamberg, EMBO J. 4 (1985) 3079^ 3085. [13] M. Nakao, D.C. Gadsby, Nature 323 (1986) 628^630. [14] D.W. Hilgemann, Science 263 (1994) 1429^1432. [15] R.F. Rakowski, J. Gen. Physiol. 101 (1993) 117^144. [16] M. Holmgren, R.F. Rakowski, Biophys. J. 66 (1994) 912^ 922.
[17] X. Wang, F. Jaisser, J.-D. Horisberger, J. Physiol. 491 (1996) 579^594. [18] U. Hasler, X. Wang, G. Crambert, P. Beguin, F. Jaisser, J.-D. Horisberger, K. Geering, J. Biol. Chem. 273 (1998) 30826^30835. [19] K. Geering, I. Theulaz, F. Verrey, M.T. Ha«uptle, B.C. Rossier, Am. J. Physiol. 257 (1989) C851^C858. [20] J.-D. Horisberger, P. Jaunin, P.J. Good, B.C. Rossier, K. Geering, Proc. Natl. Acad. Sci. USA 88 (1991) 8397^8400. [21] P. Jaunin, J.-D. Horisberger, K. Richter, P.J. Good, B.C. Rossier, K. Geering, J. Biol. Chem. 267 (1992) 577^585. [22] M. Taglialatela, L. Toro, E. Stefani, Biophys. J. 61 (1992) 78^82. [23] A.C.S. Costa, J.W. Patrick, J.A. Dani, Biophys. J. 67 (1994) 395^401. [24] A. Efthymiadis, J. Rettinger, W. Schwarz, Cell Biol. Int. 17 (1993) 1107^1116. [25] X. Wang, J.-D. Horisberger, Am. J. Physiol. Cell Physiol. 37 (1995) C590^C595. [26] P.J. Garrahan, I.M. Glynn, J. Physiol. 192 (1967) 189^216. [27] A. Bahinski, M. Nakao, D.C. Gadsby, Proc. Natl. Acad. Sci. USA 85 (1988) 3412^3416. [28] S. Heyse, I. Wuddel, H.-J. Apell, W. Stu«rmer, J. Gen. Physiol. 104 (1994) 197^240. [29] R.F. Rakowski, F. Bezanilla, P. De Weer, D.C. Gadsby, M. Holmgren, J. Wagg, Ann. NY Acad. Sci. 834 (1997) 231^ 243. [30] R.D. Pelu¡o, J.R. Berlin, J. Physiol. 501 (1997) 33^40. [31] G. Blanco, A.W. DeTomaso, J. Koster, Z.J. Xie, R.W. Mercer, J. Biol. Chem. 269 (1994) 23420^23425. [32] A.T. Beggah, P. Beguin, P. Jaunin, M.C. Peitsch, K. Geering, Biochemistry 32 (1993) 14117^14124. [33] R. Goldshleger, S.J.D. Karlish, A. Rephaeli, W.D. Stein, J. Physiol. 387 (1987) 331^355.
BBAMEM 77584 2-4-99